An Improved Face Recognition Technique Based on Modular Multi-directional Two-dimensional Principle Component Analysis Approach

摘要

In this paper, a new method named modularmulti-directional two-dimensional principle componentanalysis (M2D2DPCA) is proposed for face recognition. First,the original images are rotated at some predeterminedangles so that we may extract features from the images inany direction. Then we divide the rotated images intosmaller sub-images and apply 2DPCA approach to each ofthese sub-images. Finally we propose a fusion methodnamed modular multi-directional 2DPCA (M2D2DPCA) tocombine a bank of preliminary results in different directions.Compared with conventional 2DPCA based algorithms, theadvantage of the proposed method is that it can extractsignificant features from the images in any direction andavoid the effects of varying illumination and facialexpression. The results of the experiments on ORL and Yaledatasets show that the proposed M2D2DPCA method canobtain a higher recognition rate than the conventional2DPCA based methods.